I've had my latest creation up and running for about 2 weeks now but haven't gotten around to posting it until now.

Here are the specs:
-core I7 930 2.8Ghz (auto overclocks to 2.93)
-6Gb (3x2Gb sticks) triple channel corsair XMS3 DDR3 RAM
-Gigabyte GA-X58A-UD3R
-Western Digital 1TB green drive
-Saphire 5770 vapor-x
-Seasonic X-650
-LG Blue Ray drive
-Scythe Mugen 2
-Thermalright HR-03 Rev. A
-1x scythe s-flex 92mm 1500 RPM (run at 5 volts ~550 RPM)
-1x Scythe slipstream 120mm 1200 RPM ( run at 5 volts ~ 830 RPM )
-1x Scythe slipstream 120mm 1200 RPM ( automatically PWM controlled by graphics card )
-Antec Solo
-Windows 7 home premium 64-bit

Originally bought an Antec earthwatts 750 watt power supply but it was too loud for my ears.

I also hoped that the stock cooling of the vapor-x card would be sufficiently quiet and while it was pretty good at idle under load it would spin up very quickly. so rather than trying to mod the cheaper and potentially better accelero s1 I went ahead and bought a HR-03 and mounted a 12 cm slipstream to it.

After removing one of the front fans i decided to suspend the other one in the bungees, one of the many tricks I stole from Speedkar9 along with the cable routing for the HD.

The Seasonic performed great at Idle but under load the extra heat from the CPU caused the fan to ramp up a lot so I built a duct out of plexi and mounted it to the rear case fan. In doing this I knocked out the metal drive bay covers allowing more airflow for the CPU.

This is my second good experience with the Mugen 2 and I love the PWM fan which it comes with and wish i had one for the rear of the case. I have it set to run at ~320 RPM at idle and ~900 RPM at full load.

I love the amount of connectors the Gigabyte board has but do wish it had a PCI-E x16 slot be the second slot down instead of another PCI-E x1 slot to allow more room for graphics card cooling but thats a minor thing.


DSC01832


DSC01815
Here is the duct i made out of plexi.


DSC01817
It was made in two pieces to be able to install it and then screwed together with self taping screws and supported by a twist tie.


DSC01822
In this photo you can see how much clearance the PSU fan has.


DSC01826


DSC01825


DSC01821
tons of connector options from both the motherboard and graphics card.


DSC01829

Ambient 19C

Idle temps:
core
0. 34
1. 33
2. 35
3. 33
CPU 30
CPU fan 230 RPM
HD 25
GPU 30
GPU fan 32% (~400 RPM)

stress testing was done with prime 95, furmark, OCCT, and ATI tool
Load temps:
corre
0. 70
1. 70
2. 70
3. 67
CPU 56
CPU fan 920 RPM
HD 31
GPU 75
GPU fan 45% (~550 RPM)

Overall I am fairly satisfied with the amount of noise at the moment the lodest components are the Hard drive and rear case fan.

Any advice would be greatly appreciated.

*edit*
Updated temps and graphics cooling

Updated pics may be coming later.

You forgot to paint the interior black and flip the motherboard/ PSU lol.

Nice build!

I think 90C load is a bit high for a card with such a huge heatsink and two (effectively) 92mm fans.
Soft mount the rear case fan.

Ya I don't have quite the extreme modding skills you do or the courage to tear into such a nice shinny case (although when done well the result is even shinier).

I thought the same thing about the graphics card i was hoping for improved performance with the HR-03 vs stock cooling but the result, although very quiet with the fans only running a litlle over 500 RPM, is temperatures that are almost identical. I initially tried to run the card with out a fan mounted directly on it but temperatures were too high.

I am thinking about designing some sort of duct not sure how though so any help on this issue would be appreciated.

soft mounting sounds like a great idea ill have to see if i have anything lying around that could work or just buy some soft mounts.

A little on the high side, the temps, especially GPU under load. 5770 is a cool card, and HR-03 is a mighty HS, plus you've got 2 fans on it.
Now I know 4770 is not the same league, but even with fanless AC S1 I never saw temps higher than 65 :/
I don't think a duct would help here - you already have a 12 cm close to it. I guess the HR-03 is a dated design, and one meant for hi performance hi airflow situation.
Just my 2 c.

Yes, it's nice and shiny but also scratch and dust magnet. There is a simple trick to preserve its original looks: car wax A simple carnauba wax will get rid of minor scratches and preserve the shine. The sooner applied the better. I've polished mine with car wax a year after I bought it and was still able to remove hairline scratches. Now it looks brand new
Also, all surfaces become very smooth so dusting is easier.
Give it a try.

Those temps are high for the GPU... Strange. I have a 9800gtx+ (higher TDP than the 5770) with the HR-03 GT and a 92mm Nexus fan and temps are between 40/65°C (idle/max load).

I'm working on a new duct for the GPU, check out my sig for an idea I'm experimenting with.

+1, My 4850 + S1 gets 40/65 idle/load with one 92mm fan @5v soft mounted in the HD bays.

Reseat the heatsink with another TIM and see what it does to temps.

pull the side off of the case, and run the same stress tests... if the temps go down, there isn't enough air going through the case.

the video card test to use is occt, which is based on furmark, but more intense... 90c would not be too bad for occt gpu test, depending on how long you ran it for.

I would snip out the rear fan grill and soft-mount your rear exhaust fan, which might make a bit of different to the noise coming from that fan.

For GPU temps, maybe open up a few more of the PCI slots, and add a little baffle between the left side (as you are looking at it from your pictures) of the fan mounted on your GPU heatsink and the floor of the case. Hopefully, the effect of this will be that all the air that is being pushed out of the PCI slot area will be forced to go over your GPU heatsink.

To assist with this, I would seal off the grill vent to the right of the PCI slots, to prevent any way for air to flow out of the lower part of your case other than through the PCI slots (and through the GPU heatsink).

Hope that idea helps!

*edit* just realised that was I was suggested was exactly what frenchie had already tried in his Solo (linked above).

*mumbles something about great minds thinking alike*

Anyway, the last picture in frenchie's first post shows what I was talking about with the baffle thingy around the GPU heatsink.

I have tried pulling off the case door and that causes temperatures to shoot up I am experimenting with a 120mm fan on the GPU and will have some more info when i have time to run some test but so far it looks like the results are better.

I attatched a 12cm fan to my graphisc card cooler in place of the 92mm and am running off of the auto PWM control of the card this means that at idle it is at 32% and at load it ramps up to 45% because despite attempts to get an RPM signal off the fan i was unable to. However because it is PWM control instead of voltage you can do mathematical calculations to get a pretty close aproximation so that would mean that it is running at ~400 RPM at idle and ~550 RPM at load both of which are less than 5 volts temps are also much better now I retested all temps due to much colder ambient temps now.

Ambient 19C

Idle:
Core #
0. 34
1. 33
2. 35
3. 33
HD 25
CPU 30
CPU fan 230 RPM
GPU 30
GPU fan 32%

Load:
Core #
0. 70
1. 70
2. 70
3. 67
HD 31
CPU 56
CPU fan 920 RPM
GPU 75
GPU fan 45%

I am looking at making some sort of Duct/Baffle like frenchie and JamieG have suggested. I am also hoping to get some soft mounts for the rear fan and change the rear fan to a PWM model and link it to the CPU.

*edit*
I should also note that I added OCCT to my list of apps to stress the card although this did not seem to have a huge impact on temps.

Also if anyone has any ideas for getting RPM off the graphics controlled fan other than running the RPM wire to the mobo (this does not work) please let me know.

Did you run the black ground wire and the RPM (yellow) wire to the motherboard from the fan? It works for me in my custom fan controller.

Did you reseat the HR-03 with new TIM?

I just tried attaching the clack gound as well and it caused the fan to run full tilt with no PWM control from the graphics card.

Yes i forgot to mention that before i tried the 12cm fan I reseated the hr-03 with new TIM and even tried the thermalright TIM to see if it worked any better than AS5 this did not seem to have any affect so i switched back to AS5 non of this made any real difference in temps.

If you have any further ideas for RPM monitoring please let me know.

Hi, on my rig I've had no problem reading fan RPMs from the graphics and PSU fans by routing just the rpm (usually yellow) wire to the motherboard header.
In both cases these are PWM fans and controlled by the videocard / PSU. Arctic Cooling PWM F9 and F8 fans, the GPU programed to drop to 25% ~600rpm and the PSU does what it wants, cold idle is I think ~500rpm and below the speed the fan puts out a reliable fan signal. Under heavy load it climbs 700~900rpm where the speed signal stabilises.
I actually have 2 fans on the video card (GTX260) and one send tach signal to the video cards header, readable with GPU-Z and the other goes to a motherboard header and visable in Speedfan.
See my thread here.

Maybe you can test your fan(s) just on the motherboard header and see if they produce a speed signal and down to what speed.

Regards, Seb

I have finally figured out what is caussing my fan not to report RPM signals.
When I set the fan to 100-65% it gives a RPM reading however anything below this delivers no RPM signal and at these speeds it is far too loud. The other thing i noticed is the fan is making an audible ticking that was not present before hooking it up to the cards PWM controller. This means the card is delivering a straight unfiltered PWM square wave causing the motor to audibly kick each time it is put in the on position.
All pc fans have 2 pickups per rotation to calculate RPM but these only function if the fan is being powered as these pickups pass by with PWM the fan is only powered part of the time.
So what is necessary to solve both the RPM feedback and the ticking noise is a capacitor to smooth th PWM signal as explained in the Voltage and Power control section of this page. http://pcbheaven.com/wikipages/PWM_Modulation/?p=0#5 The problem is I am not sure what capacitor to use I believe speedkar9 has some experience with this when he built his PWM fan controller for his SOLO. If anyone else has any ideas please let me know.

I can't help but reply here as an electronic technologist.

Your theory seems correct, just to add:
The power from the graphics card is a PWM, meaning that the fan's power is being pulsed and is not constant, this jerks the fan, making that clicking noise. This will definitely kill any RPM signal.

The reason why it's working at higher speeds is that there may be some capacitance in your fan so the average voltage is higher, allowing constant power. We need to add a capacitor in parallel to the fan to emulate that.

When I played around with PWM controllers, I usually found the size by trial and error.
I currently have two fans in a parallel config feeding off a voltage controlled pwm. My frequency of operation is about 10- 20khz or so, yours may be different. I used a 470uF electrolytic capacitor and it seems to work alright.
Connect it like this:



MAKE SURE YOU PUT THE CAP IN THE RIGHT POLARITY
or else it will blow up, i've experienced it.

If you're lazy, stop here.
---------------------------------------------
The question is adding a resistor, whenever the capacitor is fully discharged, it acts like a short. And whenever the card applies power to it, the power leads get shorted. I don't have a resistor with my capacitor because my homemade circuit has a voltage regulator/buffer in there, but if you have a few laying around, you can try putting it in series with the capacitor.

The value of that will change your capacitor value.
Using the formula: F=1/RC, we can solve for C if we have F and R.
Say you have 100 ohm resistors lying around, and your circuit produces 20khz. C will be 500nF, multiply that by 10 for a longer discharge you get 1uF.
I haven't tried this resistor method but according to theory, it should work.




If you still hear the ticking noise, use a larger cap or resistor (if used), but keep in mind the larger the discharge time, the more time the fan will take to react to the input signal.

I have gotten a PWM splitter cable and a slipstream PWM fan for the rear case fan as well as soft mounts. I will be doing an update with pics sometime soon.

I have also installed a baffle for the graphics card as some have suggested.

I tried installing a 750uF capacitor both with and without a resistor with no luck i then tried upping the capacitor value to 1000uF and it seems to change nothing the speeds at which it reads RPM are the same i would have expected some sort of decrease in RPM readable speeds but have noticed no difference. not sure what the issue is i have installed the capacitor in parallel with the fan as recommended. Any ideas?

Looking forward to the pics and results of that !

It is actually quite disappointing after further experimentation I realized that I can drop my graphics card temps by about 9 degrees C if i remove the baffle I guess there was just not enough room to move all the air through that small hole this also might have had more of an effect if I was passively cooling the card.

Nice build you got there. You mention the gpu temps droping after removing the baffle, are you referring to the plexi one around the cpu? It does make sense, while the gpu cooler receives enough airflow, there is not a good exhaust path (except for that removed pci slot cover). After removing the cpu duct, the case exhaust fan can remove the heat produced by the gpu more effectively..

As for the issue with the rpm signal, one alternative explanation is that the motherboard/gpu, whichever is reading the signal, simply can`t report rpm values below a certain threshold (800rpm for example) something I`ve experienced before a few times. Either way, using a pwm fan on the gpu cooler might be more appropriate and could eliminate the clicking issue (though not all pwm fans are immune to clicking at low speeds).

The baffle which i am reffering to is not pictured it is the one that frenchie and some others recomended above. It is a piece of plexi positioned vertically underneath the GPU cooler at the back to try to make all of the air go through the cooler. My guess is it was too restrictive and did not let enough air out.

I have also removed all of the PCI slot covers which has improved temps slightly. I have however tried to make sure there is good front to back airflow (no air coming in through the PCI slots) with the fron 92mm fan.

The CPU/ PSU duct had little effect on the GPU it did raise the CPU temps slightly but made the PSU much quieter which was the goal.

AS for the fan I know the motherboard can read down to at least 500 RPM (much lower than the minimum of the GPU fan 1000 RFPM) and I have read 500 RPM with that very same header on the mobo. The clicking is farily quiet and can only be heard with the case door open and ear next to the fan. I think i will just accept that there will be no RPM feedback from that fan.

Nicely done! I have the same case and am completely in love with it 7 months later. So freaking quiet I can barely stand it

the key is my SSD. Could never go back. Ever.

too bad most just have no idea how cool a quiet computer is.

Haha my friends are already. They always think that PC is always quiet, but i started to irritate myself because of the HDD noise.

If you want quiet, an SSD will be your best friend. really. It is amazing.